SIMBA FEATURE NUMBER MISMATCH ERROR

[13281306705]

Hello, I'm sorry to bother you. When I clicked on 'Run Model' under the 'Run machine model' option in simba, I encountered an error: SIMBA FEATURE NUMBER MISMATCH ERROR: Mismatch in the number of features in input file D:/simbaAndDeepLabCut/123/loxy/models/generated_ Models/run. sav, and what is expected by the model run The model expects 153 features The data contains 155 features. Do you know what caused this?

[sronilsson]

Hi @13281306705! Do you have an error msg from the operating system terminal?

This error happens when you build a model using the data files inside the project_folder/csv/targets_inserted directory, and each file in this directory contains 153 features (all columns minus the annotation columns and the body-part columns).

You then try to run the model on your data inside the project_folder/csv/machine_results directory, and SimBA finds 155 columns in a file. Do you see two columns that may have sneaked in by mistake in a project_folder/csv/machine_results file?

Note: The error msg reads, SIMBA FEATURE NUMBER MISMATCH ERROR: Mismatch in the number of features in input file D:/simbaAndDeepLabCut/123/loxy/models/generated_ Models/run. sav It is a error msg bug, which I will fix now, and shouldn't affect the program or cause the error. It should read *SIMBA FEATURE NUMBER MISMATCH ERROR: Mismatch in the number of features in input file TheVideoFileNameYourTrying toAnalyze

[sronilsson]

... is it possible you selected a file to use in validation that lives in project_folder/csv/targets_inserted rather than in project_folder/csv/features_extracted ?

[13281306705]

... is it possible you selected a file to use in validation that lives in project_folder/csv/targets_inserted rather than in project_folder/csv/features_extracted ?…您是否选择了一个位于 project_folder/csv/targets_inserted 而非 project_folder/csv/features_extracted 中的文件用于验证？

Hello, this is the detailed error message: During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "d:\anaconda3\envs\simba\lib\tkinter__init.py", line 1699, in call return self.func(*args) File "d:\anaconda3\envs\simba\lib\site-packages\simba\SimBA.py", line 372, in button_runvalidmodel = Button(label_model_validation, text='RUN MODEL', fg='blue', command=lambda: self.validate_model_first_step()) File "d:\anaconda3\envs\simba\lib\site-packages\simba\SimBA.py", line 566, in validate_model_first_step clf_path=self.modelfile.file_path) File "d:\anaconda3\envs\simba\lib\site-packages\simba\model\inference_validation.py", line 53, in init__ output_df[probability_col_name] = self.clf_predict_proba(clf=clf, x_df=data_df, model_name=classifier_name, data_path=clf_path) File "d:\anaconda3\envs\simba\lib\site-packages\simba\mixins\train_model_mixin.py", line 971, in clf_predict_proba raise FeatureNumberMismatchError(f'Mismatch in the number of features in input file {data_path}, and what is expected by the model {model_name}. The model expects {str(clf.nfeatures)} features. The data contains {len(x_df.columns)} features.') simba.utils.errors.FeatureNumberMismatchError: Mismatch in the number of features in input file D:/simbaAndDeepLabCut/123/lxy/models/generated_models/run.sav, and what is expected by the model run. The model expects 153 features. The data contains 155 features.

After listening to your description, I roughly understand the reason for the error. Indeed, I chose project folder/csv/targets Inserted for validation.Thank you very much for your answer.

[sronilsson]

Got it! Yes, its likely that those files contain two additional columns (your behavior annotations) which we don't want when running validation.